Synthetic Model Complex of the Key Intermediate in Cytochrome P450

Jan 9, 2019 - Despite much research, the exact nature of the critical “Intermediate I” responsible for the key N–N coupling step in P450nor is u...
7 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Synthetic Model Complex of the Key Intermediate in Cytochrome P450 Nitric Oxide Reductase Ashley B. McQuarters,† Elizabeth J. Blaesi,‡ Jeff W. Kampf,† E. Ercan Alp,§ Jiyong Zhao,§ Michael Hu,§ Carsten Krebs,‡ and Nicolai Lehnert*,† †

Department of Chemistry and Department of Biophysics, University of Michigan, Ann Arbor, Michigan 48109, United States Department of Chemistry and Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania 16802, United States § Advanced Photon Source (APS), Argonne National Laboratory (ANL), Argonne, Illinois 60439, United States Downloaded via UNIV OF NEW ENGLAND on January 9, 2019 at 16:08:38 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: Fungal denitrification plays a crucial role in the nitrogen cycle and contributes to the total N2O emission from agricultural soils. Here, cytochrome P450 NO reductase (P450nor) reduces two NO to N2O using a single heme site. Despite much research, the exact nature of the critical “Intermediate I” responsible for the key N−N coupling step in P450nor is unknown. This species likely corresponds to a Fe-NHOH-type intermediate with an unknown electronic structure. Here we report a new strategy to generate a model system for this intermediate, starting from the iron(III) methylhydroxylamide complex [Fe(3,5-Me-BAFP)(NHOMe)] (1), which was fully characterized by 1H NMR, UV−vis, electron paramagnetic resonance, and vibrational spectroscopy (rRaman and NRVS). Our data show that 1 is a high-spin ferric complex with an Nbound hydroxylamide ligand that is strongly coordinated (Fe−N distance, 1.918 Å; Fe-NHOMe stretch, 558 cm−1). Simple one-electron oxidation of 1 at −80 °C then cleanly generates the first model system for Intermediate I, [Fe(3,5-MeBAFP)(NHOMe)]+ (1+). UV−vis, resonance Raman, and Mössbauer spectroscopies, in comparison to the chloro analogue [Fe(3,5-Me-BAFP)(Cl)]+, demonstrate that 1+ is best described as an FeIII-(NHOMe)• complex with a bound NHOMe radical. Further reactivity studies show that 1+ is highly reactive toward NO, a reaction that likely proceeds via N−N bond formation, following a radical−radical-type coupling mechanism. Our results therefore provide experimental evidence, for the first time, that an FeIII-(NHOMe)• electronic structure is indeed a reasonable electronic description for Intermediate I and that this electronic structure is advantageous for P450nor catalysis because it can greatly facilitate N−N bond formation and, ultimately, N2O generation.

■

up by denitrifying bacteria and fungi, which reduce NO3− stepwise to nitrous oxide (N2O; nar = nitrate reductase; nir = nitrite reductase; nor = nitric oxide reductase):4,5

INTRODUCTION Nitrogen is an essential building block for all forms of life on Earth.1−3 The primary sources of nitrogen for plants are ammonia (NH3/NH4+) and nitrate (NO3−; from minerals). Ammonia is generated from dinitrogen (N2) through biological (nitrogenases) and synthetic (the Haber−Bosch process) nitrogen fixation and then utilized by plants. However, plants only use a fraction of the ammonia that is provided by fertilization of agricultural soils, and the remainder is metabolized by large microbial communities that live in the soil and seawater. Ammonia is oxidized by nitrifying microbes to nitrite (NO2−) and nitrate (NO3−), which are then picked © XXXX American Chemical Society

nar

nir

nor

NO3− ⎯→ ⎯ NO2− → • NO ⎯→ ⎯ N2O

(1)

In the final step of denitrification (not shown above), N2O is further reduced to N2 (by denitrifying bacteria only) to close the nitrogen cycle. Overfertilization of agricultural soil has Received: October 17, 2018

A

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

1851 and 530 cm−1, respectively.18 This is followed by a direct hydride transfer from NAD(P)H to the heme-{FeNO}6 complex.11 This reaction causes a characteristic change in the Soret band, which shifts from 431 to 444 nm.17 The lifetime of the resulting “Intermediate I” is only 100 ms, making it difficult to further study this species. From resonance Raman (rRaman) spectroscopy, the Fe−NO stretching frequency of this intermediate is 596 cm−1.19 On the basis of the above reaction sequence, Intermediate I corresponds to a nitroxyl level complex of the type {FeN(H)nO}8 (n = 0−2) of an unknown protonation state. Density functional theory (DFT) and quantum mechanics/molecular mechanics (QM/ MM) calculations have helped to shed further light on the details of the reaction.20,21 From DFT, the {FeNO}6 complex is best described as an FeII-NO+ species, which is consistent with its spectroscopic properties (see above). Direct hydride transfer from NAD(P)H initially generates a ferrous hemeHNO species, as shown in Scheme 1, which is very likely Nprotonated [based on (a) total energy considerations from DFT 20 and (b) the experimental findings for the myoglobin(II)-NHO adduct22]. At this point, the ferrous heme-NHO species could react directly with a second equiv of NO to form the N−N bond and generate a ferric hyponitrite intermediate. Alternatively, because of the strong donicity of the cysteinate ligand, the ferrous HNO complex could be further protonated, resulting in a formally FeIV-NHOH− species.20 On the basis of recent QM/MM calculations, the electronic structure of this intermediate could also be described as an FeIII-(NHOH)• complex, where a NHOH radical is bound to a low-spin (ls) iron(III) and the spins are antiferromagnetically coupled (total spin S = 0).21 Then, the subsequent addition of NO generates again a ferric hyponitrite complex, which readily decomposes into N2O and H2O, closing the catalytic cycle (see Scheme 1). On the basis of these results, the critical Intermediate I is identified with either the ferrous heme-NHO complex or the corresponding doubly protonated species. There is some further experimental evidence that suggests that Intermediate I might be the doubly protonated species. Pulsed radiolysis of hydroxylamine (NH2OH) was shown to generate the (NHOH)• radical in water.23 The irradiation of NH2OH in the presence of ferric P450nor forms a species with a UV−vis spectrum identical with that of Intermediate I (Soret band at 444 nm).11 This result implies that Intermediate I is an NHOH-type complex. Further experimental information about this species comes from recent magnetic circular dichroism (MCD) and Mössbauer studies.24 First, the MCD data of Intermediate I show no paramagnetic MCD C-term signal at the Soret band position of this species (445 nm, ∼22500 cm−1), which provides direct evidence that Intermediate I is diamagnetic. 25 These results were confirmed by 57 Fe Mössbauer spectroscopy. Interestingly, the isomer shift (δ) determined for Intermediate I is within the range for ls ferric hemes (δ ≈ 0.15−0.25 mm/s). Unfortunately, no Mössbauer data are available for ferrous heme-NHO complexes for comparison. DFT-calculated isomer shifts are similar for the ferrous heme-NHO (δ = 0.26 mm/s) and FeIII-(NHOH)• (δ = 0.23 mm/s) complexes, so the protonation state of Intermediate I cannot be assigned based on these results. However, the DFT calculations do not favor the FeIV-NHOH− valence tautomer (predicted δ = 0.09 mm/s). In summary, the protonation state and electronic structure of the key intermediate in P450nor catalysis, Intermediate I, are still not

become a common practice in the developed world (to ensure that nitrogen is never a limiting nutrient for crop growth), which leads to an overabundance of bioavailable nitrogen.6,7 As a result, denitrifying bacteria and fungi metabolize increasing amounts of NO3−/NO2− and release steadily increasing quantities of the gaseous intermediates NO and N2O into the atmosphere. This is a major environmental concern because both NO and N2O deplete the ozone layer and N2O is also a potent greenhouse gas.7 These environmental impacts have led to great interest in the enzymes involved in nitrification and denitrification and their catalytic mechanisms.8 One enzyme involved in the two-electron reduction of NO to N2O that prevents accumulation of this toxic metabolite is cytochrome P450 nitric oxide reductase (P450nor).9,10 This enzyme is found in soil-dwelling, denitrifying fungi like the ubiquitous Fusarium oxysporum but also yeasts (for example, Trichosporon cutaneum). P450nor belongs to the cytochrome P450 superfamily (Cyt. P450s), which are enzymes that mostly perform oxidative (especially monooxygenase) chemistry. Unlike the typical Cyt. P450s, P450nor is unique for its function as a reductase, even though the active site of P450nor shares strong structural similarities with its monooxygenase counterparts, in particular the single heme b with a proximal (axial) cysteinate ligand.11,12 Figure 1 shows the crystal

Figure 1. NO-bound ferric heme in the active site of P450nor from F. oxysporum. This image was generated using PyMol from PDB code 1CL6 (adapted from ref 13).

structure of the NO-bound ferric heme complex of F. oxysporum P450nor.13 This enzyme performs NO reduction following eq 2 with a turnover frequency of ∼1200 s−1. The necessary electrons are derived from NAD(P)H without the aid of an electron-transfer protein.14 2NO + NAD(P)H + H+ → N2O + H 2O + NAD(P)+ (2)

In the first step of catalysis, the ferric heme complex of P450nor binds NO to form a ferric heme nitrosyl, or {FeNO}6 complex in the Enemark−Feltham notation (where the superscript “6” represents the number of Fe(d) electrons plus the NO(π*) electrons),15 as the first intermediate of the reaction.16,17 NO binding goes along with a shift of the Soret band from 413 to 431 nm.17 The NO complex is further characterized by N−O and Fe−NO stretching frequencies of B

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 1. Proposed Mechanism for N2O Production by P450nora

a

The oval represents a generic porphyrin ligand.

resolved. Therefore, it is the goal of this work to use simple synthetic models to further elucidate the nature, electronic structure, and reactivity of this species. In this work, we devised a convenient route to synthesize an FeIII-(NHOR)•/FeIV-NHOR−-type model complex for Intermediate I through one-electron oxidation of a corresponding FeIII-NHOR− precursor, as shown in Scheme 2. We first

Scheme 3. Target Complex 1, Which Serves as a Precursor for the Proposed FeIII-(NHOH)•/FeIV-NHOH− Intermediate in the Catalytic Cycle of P450nora

Scheme 2. Synthetic Route Used To Generate a Model for Intermediate I of P450nor

prepared the precursor [Fe(3,5-Me-BAFP)(NHOMe)] (1) and fully characterized this complex by UV−vis, 1H NMR, and electron paramagnetic resonance (EPR) spectroscopy, and Xray crystallography (which represents the first crystal structure of a ferric heme hydroxylamide complex reported to date). Because of the fact that iron(IV) porphyrins are known to be highly reactive, a bis(picket fence) porphyrin, 3,5-Me-BAFP2− (3,5-Me-BAFP 2− = tetra[2,6-bis(3,5-dimethylphenoxy)phenyl]porphyrin dianion), was used here to protect the bound hydroxylamide ligand (see Scheme 3). The oneelectron oxidation of the FeIII-NHOMe− precursor via chemical oxidation results in the target complex [Fe(3,5-MeBAFP)(NHOMe)]+ (1+), which is a metastable species at −80 °C. We then investigated the spectroscopic properties of this species, using EPR, rRaman, and Mössbauer spectroscopies, which ultimately shows that this intermediate is best described as an FeIII-(NHOMe)• complex and not the FeIV valence tautomer. This species can be generated in ∼85% yield in solution. Finally, the reactivity of this intermediate with NO was investigated, which leads to the fast generation of a ferric product complex. The implications of these results for the mechanism of P450nor are finally discussed. Note that, compared to the P450nor active site, our model system lacks the axial thiolate ligand, which might further influence the reactivity of Intermediate I in the enzyme. Similar to the

a

See also Figure 3.

protein matrix, the sterically demanding pickets used in our model provide shielding for reactive intermediates that are coordinated to the heme.

■

EXPERIMENTAL PROCEDURES

All reactions were performed under inert conditions using Schlenk techniques. The preparation and handling of air-sensitive materials was carried out under a dinitrogen atmosphere in an MBraun glovebox equipped with a circulating purifier (O2, H2O < 0.1 ppm). Nitric oxide (Cryogenic Gases Inc., 99.5%) was purified by passage through an ascarite II column (NaOH on silica), followed by a cold trap at −80 °C to remove higher-order nitrogen oxide impurities. All solvents (including deuterated solvents) and 1-methylimidazole were distilled from CaH2 under dinitrogen and then degassed via five freeze−pump−thaw cycles. Tetrabutylammonium hexafluorophosphate was recrystallized from ethanol. The purified solvents were stored over appropriately sized activated molecular sieves in the glovebox until used. 1,1′-Diacetylferrocene was purchased from Fisher Scientific, tris(4-bromophenyl)ammoniumyl hexachloroantimonate was purchased from Sigma-Aldrich, and O-methylhydroxylamine hydrochloride was purchased from TCI America. All of these materials were used without further purification. Deprotonation of C

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

The first three terms represent the electronic Zeeman effect and zerofield splitting, the fourth term describes the interaction between the nuclear quadrupole moment and electric field gradient, the fifth term represents the magnetic hyperfine interaction of the electronic spin with the 57Fe nucleus, and the last term describes the 57Fe nuclear Zeeman effect. Crystal Structure Determination of 1. A brown plate of 1 of dimensions 0.22 × 0.15 × 0.02 mm was mounted on a Rigaku AFC10K Saturn 944+ CCD-based X-ray diffractometer equipped with a low-temperature device and a Micromax-007HF copper-target microfocus rotating anode (λ = 1.54187 Å), operated at 1.2 kW power (40 kV, 30 mA). The X-ray intensities were measured at 85 K, with the detector placed at a distance of 42.00 mm from the crystal. A total of 3461 images were collected with an oscillation width of 1.0° in w. The exposure time was 5 s for the low-angle images and 20 s for the high-angle images. Rigaku d*trek images were exported to CryAlisPro34 and corrected for absorption. The integration of the data yielded a total of 61858 reflections to a maximum 2θ value of 147.92°, of which 9532 were independent and 8139 were greater than 2σ(I). The final cell constants in Table S2 are based on the xyz centroids of 17282 reflections above 10σ(I). Analysis of the data showed negligible decay during data collection; the data were processed with CrystalClear 2.035 and corrected for absorption. The structure was solved and refined with the Bruker SHELXTL (version 2016/6) software package,36 using the space group P1̅ with Z = 1 for the formula C110H100N6O10Fe. All non-hydrogen atoms were refined anisotropically, with the hydrogen atoms placed in idealized positions. Full matrix least-squares refinement based on F2 converged at R1 = 0.0957 and wR2 = 0.2816 [based on I > 2σ(I)] and R1 = 0.1028 and wR2 = 0.2938 for all data. DFT Calculations. The structure of [Fe(P)(NHOMe)] (P2− = porphine dianion) was fully optimized for the sextet (S = 5/2) ground state of complex 1, using the BP86 functional37,38 and the TZVP basis set.39,40 Vibrational frequencies calculated for the optimized structure show no imaginary frequencies. In all calculations, convergence was reached when the relative change in the density matrix between subsequent iterations was less than 1 × 10−8. All of these calculations were performed using the program package Gaussian 09.41 Preparation of Salt-Free O-Methylhydroxylamine. A total of 11.2 g of O-methylhydroxylamine hydrochloride (0.134 mmol) was dissolved in 95 mL of a sodium hydroxide solution (5 M). Then, the product O-methylhydroxylamine was distilled off under dinitrogen as a clear oil. Yield: 4.53 g (0.096 mol, 72%). 1H NMR (CDCl3, 400 MHz): δ: 3.554 (s, 3H), 5.437 (b, 2H). Synthesis of [Fe(3,5-Me-BAFP)(NHOMe)] (1). Under an inert atmosphere, 244 mg of [Fe(3,5-Me-BAFP)(PF6)] (0.137 mmol) was dissolved in 5 mL of dry 2-methyltetrahydrofuran (2-Me-THF; the reaction also works in toluene). Then, 22 mg of Li[NHOMe] (0.75 mmol) was added to the solution, and the reaction was allowed to stir until it looked complete by UV−vis spectroscopy. On the basis of UV−vis conversion, more ligand can be added (typically up to an additional ∼5 equiv) because of the poor solubility of the lithium salt. Typical reaction times are about 24 h. The resulting solution was filtered through a 0.2 μM PTFE filter, layered with hexanes, and allowed to precipitate in the −33 °C freezer. After 2 days, the reaction was filtered through a frit to give a dark-purple powder. Yield: 185 mg (0.110 mmol, 81%). UV−vis (2-Me-THF): 421, 582, 631 nm. UV− vis (toluene): 424, 582, 632 nm. 1H NMR (CD2Cl2, 400 MHz): δ −0.887 (b), −3.77 (b), 1.843 (b), 6.172 (b), 6.539 (b), 7.558 (b), 10.285 (b), 11.074 (b), 79.41 (b). Satisfactory elemental analysis required the inclusion of a solvent-derived 2-Me-THF molecule. The reported analysis is for 1·2-Me-THF. Anal. Calcd for C114H106FeN5O: C, 77.71; H, 6.06; N, 3.97. Found: C, 77.05; H, 5.87; N, 3.21. The same procedure was followed when using the precursor [Fe(3,5-Me-BAFP)(X)] with X = ClO4− or SbF6−. Brown plates suitable for X-ray diffraction were grown by layering a concentrated 2-Me-THF solution of 1 in a 5-mm-diameter glass tube with pentane at room temperature for 3 days. Chemical Oxidation of [Fe(3,5-Me-BAFP)(Cl)]. Under an inert atmosphere, 7.2 mg of [Fe(3,5-Me-BAFP)(Cl)] (0.0043 mmol) was

O-methylhydroxylamine to the corresponding lithium salt, Li[NHOMe], was carried out as previously reported.26 [Fe(3,5-MeBAFP)(Cl)],27 [Fe(3,5-Me-BAFP)(X)]28 where X = PF6− or SbF6−, [Co(TPP)],29 and [DAcFc][SbF6]30 were synthesized as previously reported. 57Fe complexes were synthesized in the same way as the natural abundance complexes, using 57FeCl2 dimethanol salt as the iron source. In this work, NO gas was added to reactions via a syringe from a bomb flask or in a stoichiometric manner using a NO-saturated solution. NO-saturated solutions were prepared under a dinitrogen atmosphere by equilibration of dichloromethane with NO gas in a Schlenk flask. The concentration of the NO gas in the solution was subsequently determined with [Co(TPP)]. The titration of [Co(TPP)] with NO(g)/CH2Cl2 was carried out in tetrahydrofuran (THF) and monitored by UV−vis spectroscopy (λmax of the Q band changes from 527 to 537 nm). Typical NO gas concentrations in dichloromethane prepared in this way were ∼1 mM. Physical Measurements. IR spectra were obtained from KBr disks with PerkinElmer BX and GX spectrometers at room temperature. Electronic absorption spectra were measured using an Analytic Jena Specord S600 instrument at room temperature. EPR spectra were recorded with a Bruker X-band EMX spectrometer equipped with Oxford Instruments liquid-nitrogen and -helium cryostats. EPR spectra were typically obtained on frozen solutions using 20 mW microwave power and 100 kHz field modulation with the amplitude set to 1 G. Sample concentrations were ∼0.1−2 mM. 1 H and 19F NMR spectra were recorded on Varian MR 400 MHz and NMRS 500 MHz instruments at room temperature. In situ UV−vis measurements were conducted using a Hellma quartz immersion probe with a 10 mm path length. Mass spectrometry (MS) data were collected on an Agilent 6230 time-of-flight high-performance liquid chromatography/mass spectrometer. Cyclic voltammograms were obtained with a CH instruments CHI600E electrochemical workstation using a three-component system consisting of a glassy carbon working electrode, a platinum counter electrode, and a silver wire pseudoreference electrode. All potentials were corrected to Fc+/Fc. UV−vis spectroelectrochemical (SEC) measurements were performed using a custom-built thin-layer electrochemical cell as previously described.27 All electrochemical and SEC measurements were carried out in the presence of ∼0.1 M tetrabutylammonium hexafluorophosphate. Nuclear resonance vibrational spectroscopy (NRVS) measurements were carried out as previously described31 at beamline 3-ID-XOR at the APS at ANL. This beamline provides about 2.5 × 109 photons/s in ∼1 meV bandwidth (8 cm−1) at 14.4125 keV in a 0.5 mm (vertical) × 0.5 mm (horizontal) spot. Samples were loaded into 4 × 7 × 1 mm copper cells. The final spectra represent the averages of four scans. The program Phoenix was used to convert the NRVS raw data to the vibrational density of states (VDOS).32,33 The rRaman measurements were performed using the 413.13 nm excitation line from a Kr+-ion laser (Spectra Physics Beam Lok 2060-RS). Raman spectra were recorded at 77 K using an Acton two-stage TriVista 555 monochromator connected to a liquidnitrogen-cooled CCD camera (Princeton Instruments Spec10:400B/LN). The total exposure time of the samples to the laser radiation was 3 min, using 1−2 accumulations, and typical laser powers were in the 17−30 mW range. Mössbauer Spectroscopy. Mössbauer data were recorded on a spectrometer from WEB Research, equipped with a Janis SVT-400 variable-temperature cryostat. All isomer shifts are quoted relative to the centroid of the spectrum of α-Fe at room temperature. Simulation of the Mössbauer spectra was conducted with the WMOSS spectral analysis package, using the spin Hamiltonian shown in the following equation: i S(S + 1) yz zz + E(Sx 2 − Sy 2) H = βS·g ·B + DjjjjSz 2 − z 3 k { Ä Å ÑÉÑ eQVzz ÅÅÅ 2 I(I + 1) η Ñ + + (Ix 2 − Iy 2)ÑÑÑ + S·A·I ÅÅÅIz − ÑÑÖ 4 ÅÇ 3 3 − gnβnB·I

D

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 2. Left: UV−vis spectra of [Fe(3,5-Me-BAFP)(PF6)] (black) and of the isolated product 1 (blue) in 2-Me-THF at room temperature. Right: EPR spectrum of 1 in toluene at 6 K. dissolved in 2 mL of dry dichloromethane. The solution was added to 5.6 mg of solid [N(C6H4Br-4)3][SbCl6] (0.0068 mmol) and agitated until all of the solid had dissolved, forming [Fe(3,5-Me-BAFP)(Cl)]SbCl6. Typical concentrations were in the 1−2 mM range. UV−vis (CH2Cl2): 373, 411, 568, 634 nm. 1H NMR (CD2Cl2, 500 MHz): δ −10.22 (s), −11.00 (s), 2.34 (b), 7.97 (b), 8.49 (b), 9.11 (b), 15.64 (s), 83.08 (b). Chemical Oxidation of 1. Typical concentrations of 1 used for these experiments were from ∼5 μM to 0.3 mM, and ∼1−1.5 equiv of the oxidant [DAcFc][SbF 6] was employed. Under an inert atmosphere, a stirring, dry toluene solution of 1 in a custom-designed flask fitted with a UV−vis immersion probe (without its metal casing) was cooled to −80 °C in a dry ice/hexanes bath. The syringes, needles, and oxidant solution were prechilled in dry ice. The oxidant [DAcFc][SbF6] was dissolved in cold dimethoxyethane (DME; typical concentrations were ∼1 mM) and slowly added to the toluene solution of 1. Correspondingly, DME reached up to 20% of the total volume of the solution at the highest concentrations of 1 employed. Completion of the reaction to yield the oxidized species 1+ was monitored by in situ UV−vis spectroscopy. Then, this solution was transferred to precooled quartz tubes (in a Dewar with a dry ice/ hexanes bath) for EPR and rRaman measurements. For preparation of the samples for Mössbauer spectroscopy, the plastic cup sample holders were cooled inside a custom-designed aluminum block that contained dry ice. N2O Detection. Under an inert atmosphere, a stirring, dry toluene solution of 1 in a custom-designed airtight flask fitted with an in situ UV−vis immersion probe (without its metal cover) was cooled to −80 °C in a dry ice/hexanes bath. After oxidation of the complex, NO gas (from a bomb flask; 2−5 mL) was added to a solution of 1+ at −80 °C via a syringe. The reaction progress was monitored using in situ UV−vis spectroscopy, applying the kinetics program of the Analytic Jena software package. Once the spectral features of the oxidized species [Fe(3.5-Me-BAFP)(NHOMe)]+ had disappeared (typical reaction times are less than 1 min), the solution was warmed up to room temperature, which took about 30 min. At this point, the UV−vis spectrum looked like that of the six-coordinate (6C) ferric heme nitrosyl complex. The headspace of the flask was then transferred via cannula into a gas IR cell and tested for N2O.

to use a derivative where the oxygen atom of the NH2OH ligand is bound to a methyl group (NH2OMe), leading to a room-temperature-stable hydroxylamine derivative. This approach has the additional advantage that the methyl substituent will help direct the nitrogen atom to bind to the iron center, generating the target complex [Fe(3,5-MeBAFP)(NHOMe)] (1), as illustrated in Scheme 3. We further decided to use the sterically shielding bis-picket fence porphyrin H2[3,5-Me-BAFP] as the ligand to stabilize the bound hydroxylamide group. The simplest reaction sequence to generate the target complex 1 would be the direct reaction of a ferric heme precursor with free NH2OMe. However, previous work by our laboratory has shown that, although NH2OR-type ligands are stable at room temperature, the reaction of free NH2OR with iron(III) porphyrins still leads to disproportionation reactions.28 We therefore speculated that akylhydroxylamide (NHOR−) salts could be reacted with ferric [Fe(Porph)(X)] precursors (where Porph2− = dianion of a generic porphyrin coligand; X− = weakly coordinating anion such as SbF6−, PF6−, ClO4−, etc.) under dry conditions to yield the desired FeIII-NHOR porphyrin complexes, as an alternative synthetic route. In order to test this idea, the hydroxylamide salt Li[NHOMe] was first prepared by direct deprotonation of NH2OMe with methyllithium in dry diethyl ether at −80 °C, as previously reported.26 The ferric hexafluorophosphate complex [Fe(3,5-Me-BAFP)(PF6)], and related complexes with weakly coordinating counterions, were prepared as previously reported.28 Then, excess Li[NHOMe] was reacted with the ferric [Fe(3,5-Me-BAFP)(X)] complexes in 2-MeTHF at room temperature to form the corresponding ferric methylhydroxylamide complex 1 in 81% yield. The UV−vis spectrum of the isolated product 1 (Figure 2) shows a broadening and a shift in the position of the Soret band from 401 to 422 nm, and a Q band shift from 525 to 582 nm. The UV−vis properties of 1 are indicative of an N-bound ferric reaction product such as [Fe(TPP)(4-MI)] (λmax = 416, 576, and 621 nm in toluene; 4-MI = 4-methylimidazolate).47 To determine whether the ferric hydroxylamide complex is highspin (hs) or low-spin (ls), EPR spectroscopy was employed. The EPR spectrum of 1 in toluene at 6 K, shown in Figure 2, exhibits effective g values, geff,x ≈ geff,y ≈ 5.94, and geff,z= 2.0, which emanate from the “Ms = ±1/2” doublet of the S = 5/2

■

RESULTS AND ANALYSIS Synthesis and Characterization of 1. Surprisingly, there are no reports of stable FeIII-NHOH-type complexes, which is likely due to the thermal instability of NH2OH itself42,43 and its propensity to disproportionate when reacted with metal centers.44−46 To increase the stability of NH2OH, we decided E

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry ground state with a rhombicity E/D of ≈0. In addition, the shifts of the β-pyrrole 1H NMR signals of the porphyrin coligand are correlated with the oxidation and spin state of the iron center.48−50 Our complex, 1, exhibits a broad signal at ∼79 ppm in the 1H NMR spectrum that is representative of a hs S = 5/2 complex. The remaining porphyrin protons are found in the 11−1.8 ppm range. Interestingly, we also observed two broad signals from the bound NHOMe− ligand at −0.89 and −3.76 ppm, as shown in Figure S3. These signals are shifted upfield, respectively, because of the ring current of the porphyrin coligand. To further confirm that the coordination of the methylhydroxylamide ligand in our Fe III-NHOMe− complex is consistent with our expectation (see the proposed structure in Scheme 3), we obtained the crystal structure of this complex shown in Figure 3. In the structural model of 1, the iron is

heme plane, 5C ferric porphyrins usually show some degree of the doming distortion, which is typically analyzed by measuring the iron atom displacement from the mean heme plane (toward the axial ligand). In the crystal structure of 1, the iron atom is displaced by 0.345 Å from the mean heme plane, which is within range (although on the lower end) of the reported iron atom displacements of 0.39−0.62 Å observed for 5C hs ferric porphyrin complexes.54 This indicates that our complex is less domed compared to other 5C ferric heme complexes from the literature. Last, we characterized complex 1 by 57Fe NRVS to identify the Fe−N stretching mode of this compound. The NRVS data in Figure S4 (top) show noteworthy features at 250 and 558 cm−1. The lower-energy band at 250 cm−1 represents the eusymmetric Fe−N(porphyrin) stretching vibration, the energy of which is representative of a hs ferric complex.55−57 By comparison to the NRVS spectrum of the precursor complex [57Fe(3,5-Me-BAFP)(SbF6)], the 558 cm−1 band in 1 could correspond to the Fe−N stretch of the methylhydroxylamide ligand. To gain further insight into the vibrational assignments, we performed DFT calculations. The structure of [Fe(P)(NHOMe)] (where P2− = porphine) was fully optimized with BP86/TZVP, and the vibrational frequencies were calculated. Overall, the optimized structure is in good agreement with our crystal structure data (Table 1). The calculated Fe−N bond length is 1.94 Å, while the crystal structure value is 1.918 Å. The calculated N−O bond length of the methylhydroxylamide ligand is 1.408 Å, whereas the crystal structure bond length is ∼0.1−0.2 Å longer (1.509−1.606 Å). The Fe−N−O bond angle of the crystal structure is 116−125°, which is consistent with the calculated bond angle of 125°. Importantly, the calculated NRVS data show very good agreement with experiment (Figure S4) and predict the Fe−N stretching frequency of the methylhydroxylamide ligand at 565 cm−1, which is very similar to the experimental band at 558 cm−1, in terms of both the energy and intensity. We therefore assign this feature to the Fe−NHOMe stretch. In summary, we have synthesized the first example of a 5C ferric hydroxylamide complex, 1, in pure form and fully characterized the complex by UV−vis, EPR, and 1H NMR spectroscopy, NRVS, and X-ray crystallography. Electrochemistry of 1. The electrochemistry of complex 1 was studied by cyclic voltammetry (CV), in order to determine whether one-electron oxidation of this complex is possible. CV of the ferric methylhydroxylamide complex exhibits a chemically reversible one-electron oxidation at a reduction potential of +0.291 V vs Fc+/Fc (in 1,2-dichloroethane), as shown in Figure 4, left. Interestingly, the one-electron oxidation product appears to be stable on the time scale for CV measurements.

Figure 3. Left: Crystal structure of 1. Hydrogen atoms and the disorder of the NHOMe− ligand are omitted for clarity. The thermal ellipsoids are shown at 30% probability. Right: Schematic drawing of 1.

disordered above and below the heme plane and the methylhydroxylamide ligand bound to the iron center is also disordered. The hydroxlyamide ligand contains a 4-fold disorder of the NHOMe− group. Importantly, the structure shows that the NHOMe− ligand is bound through its nitrogen atom as anticipated. The fact that the iron center is located above the heme plane, toward the NHOMe− ligand, is a clear indication that the complex is indeed five-coordinate (5C). The average Fe−Nporph bond length for our complex is 2.069 Å, which falls into the range of other hs (and 5C) ferric porphyrins (2.051−2.082 Å), as shown in Table 1.51−53 Interestingly, the Fe−NHOMe bond length is 1.918 Å, which is most similar to the Fe−N3 bond length in [Fe(TPP)(N3)] of 1.953 Å.51 However, our complex exhibits a longer Fe− NHOMe bond compared to the Fe−OMe bond length of 1.816 Å in [Fe(TPP)(OMe)],52 whereas it is significantly shorter than the Fe−Cl bond (2.192 Å) in [Fe(TPP)(Cl)].53 It is also worth noting that, because the iron sits above the

Table 1. Comparison of the Geometric Parameters of Selected Iron Porphyrin Complexes to 1a and the Calculated Structure of This Complex with BP86/TZVP (Using P2− = Porphine Dianion in the Calculations) complex

ΔFe−Nporhb

ΔFe−Lc

ΔO−C

ΔN−O

∠Fe−L

ref

[Fe(3,5-Me-BAFP)(NHOMe)] (1) [Fe(P)(NHOMe)], calculated [Fe(TPP)(N3)] [Fe(TPP)(OMe)] [Fe(TPP)(Cl)]

2.0695 2.10 2.066 2.082 2.051

1.918 1.940 1.953 1.816 2.192

1.487d 1.438

1.522e 1.408

125f 124 122 129

t.w. t.w. 48 49 50

1.393

Bond lengths are given in Å and the angles in degrees. bAverage value. cL denotes the axial ligand bound to the iron center. dThe NHOMe− unit is disordered, resulting in additional C−O bond lengths of 1.434, 1.447, and 1.480 Å. eAdditional N−O bond lengths = 1.509, 1.570, and 1.606 Å. f Additional ∠Fe−NH−OMe = 116, 119, and 124°. a

F

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 4. CV of a 3 mM solution of 1 in 1,2-dichloroethane that contained 0.10 M tetrabutylammonium hexafluorophosphate, using a glassy carbon working electrode, a platinum counter electrode, and a silver wire pseudoreference electrode at room temperature. Right: Schematic representation of the one-electron oxidation of 1, which could result in oxidation of the Fe-NHOMe− unit or the porphyrin co-ligand. The oval represents the porphyrin ligand.

Figure 5. Left: UV−vis spectra of the titration of a ∼ 11 μM solution of [Fe(3,5-Me-BAFP)(Cl)] (green) with the chemical oxidant [N(C6H4Br4)3][SbCl6], in dichloromethane, to form the ferric porphyin radical complex [Fe(3,5-Me-BAFP)(Cl)]+ (blue). Right: UV−vis spectra obtained upon the addition of excess oxidant, showing that the reaction is complete after the addition of 1 equiv of oxidant. The bands at 373 and 728 nm are from the oxidant [N(C6H4Br-4)3][SbCl6]. The reactions were carried out at room temperature.

Complex 1 should exhibit a (formal) FeIII/FeIV redox couple and two additional oxidations of the porphyrin macrocycle. The solvent window in 1,2-dichloroethane cuts off around +1 V, but we still observe another quasi-reversible redox event at ∼+0.90 V versus Fc+/Fc. On the basis of the CV data for complex 1 alone, it is not possible to say which group is actually oxidized at the first oxidative wave at +0.291 V. Figure 4, right, shows the three possibilities: the oxidation could either be metal-centered, transforming the complex from FeIII to FeIV, it could be hydroxylamide-centered, creating a corresponding FeIII(NHOMe)• system, or the oxidation could occur on the porphyrin coligand. The latter process would not be biologically relevant (in terms of modeling Intermediate I), and, hence, it is important to clearly rule out this possibility. In order to do this, it would be most meaningful to study a closely related complex (with the same porphyrin ligand) that actually shows a porphyrin-ligand centered oxidation and, in this way,

would allow us to determine the spectroscopic features associated with an oxidized 3,5-Me-BAFP2− porphyrin macrocycle. Oxidation of [Fe(3,5-Me-BAFP)(Cl)]. It is reported in the literature that one-electron oxidation of ferric heme chloride complexes, [Fe(Porph)(Cl)], results in the oxidation of the porphyrin coligand to yield an FeIII-(porphyrin)•+ complex (rather than oxidation of the iron center to an Fe IV species).58−61 To gain a spectroscopic handle on porphyrin ring oxidation with our 3,5-Me-BAFP2− porphyrin ligand, we carried out one-electron oxidation of [Fe(3,5-Me-BAFP)(Cl)]. The cyclic voltammogram of [Fe(3,5-Me-BAFP)(Cl)] is shown in Figure S5, which exhibits the first reversible oxidation event at an E1/2 value of +0.519 V versus Fc+/Fc in 1,2dichloroethane. Interestingly, this value is ∼200 mV more negative than the E1/2 value of [Fe(TPP)(Cl)] at +0.73 V (vs Fc+/Fc in CH2Cl2). The redox potential of our complex is actually closer to that of the complex [Fe(Tp-OCH3PP)(Cl)], G

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 6. Top left: UV−vis spectra of [Fe(3,5-Me-BAFP)(Cl)] before (black) and after the addition of ∼1 equiv of [N(C6H4Br-4)3][SbCl6] in dichloromethane (blue), leading to formation of the ferric porphyrin radical complex [Fe(3,5-Me-BAFP)(Cl)]+. The reaction product after stirring for 5.5 h under an inert atmosphere at room temperature is shown in green. Top right: Subsequent reaction of the oxidized species with excess ferrocene that re-forms the [Fe(3,5-Me-BAFP)(Cl)] complex (green). Bottom: EPR spectra of a 1.2 mM solution of the ferric chloride complex [Fe(3,5-Me-BAFP)(Cl)] and of the reaction product upon the addition of ∼1 equiv of [N(C6H4Br-4)3][SbCl6], measured at 5 K in dichloromethane.

(C 6 H 4 Br-4) 3 ][SbCl 6 ] (E 1/2 = +0.70 V vs Fc + /Fc in CH2Cl2).61 The reaction is complete upon the addition of ∼1 equiv of [N(C6H4Br-4)3][SbCl6], as shown in Figure 5, left. The addition of more than 1 equiv of oxidant simply results in an increase in the absorption bands of the oxidant itself at 373 and 728 nm (Figure 5, right). The UV−vis spectrum of the reaction product from chemical oxidation is essentially identical with that of the UV−vis SEC-generated species (Figure S6). In addition, Figure 6 shows that chemical oxidation is completely reversible upon the addition of excess ferrocene to the oxidized species [Fe(3,5-Me-BAFP)(Cl)]+. To further access the purity of the bulk material, we utilized EPR and 1H NMR spectroscopy. For these experiments, the ferric chloride precursor was dissolved in dichloromethane (1− 2 mM) and added to excess solid oxidant, [N(C6H4Br4)3][SbCl6], and the solution was then agitated until all of the solid had dissolved. The progress of the reaction was monitored by UV−vis spectroscopy. The EPR spectrum at 5 K of the ferric chloride precursor has a hs ferric signal with geff,x ≈ geff,y ≈ 5.93 and geff,z = 2.0, corresponding to the ground-state Kramers doublet of S = 5/2 with E/D of ≈0. The addition of ∼1.0 equiv of oxidant to the solution of the ferric chloride complex results in a completely EPR-silent species, shown in Figure 6, bottom. This indicates that a clean conversion to the

which contains the tetraphenylporphyrin derivative TpOCH3PP [dianion of tetra(p-methoxyphenyl)porphyrin]. This complex shows an E1/2 value of +0.64 V (vs Fc+/Fc in CH2Cl2).60 Next, we employed UV−vis spectroelectrochemistry (SEC) to determine whether or not the oxidized complex [Fe(3,5Me-BAFP)(Cl)]+ contains a porphyrin radical ligand. Oxidation of a solution of [Fe(3,5-BAFP)(Cl)] (Figure S6) results in a significant decrease in the intensity of the Soret band at 425 nm, while the Cl−-to-Fe charge-transfer feature at 373 nm remains relatively unchanged.62 This results in a broad, split Soret band. At the same time, the main Q band at 513 nm decreases in intensity as two new bands grow in at 568 and 634 nm. The oxidation is completely reversible, as shown in Figure S6. In comparison to the results for the analogous complex [Fe(TPP)(Cl)], these changes in the absorption spectrum are indeed characteristic for porphyrin ring oxidation and, in this case, formation of an FeIII-(porphyrin)•+ complex.58−60 Using a chemical oxidant, we were also able to generate the oxidized complex [Fe(3,5-Me-BAFP)(Cl)]+ in bulk at room temperature for further spectroscopic studies. For this purpose, the ferric chloride precursor complex [Fe(3,5-Me-BAFP)(Cl)] was titrated with a solution of the oxidant tris(4bromophenyl)ammoniumyl hexachloroantimonate, [NH

DOI: 10.1021/acs.inorgchem.8b02947 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry oxidized species [Fe(3,5-Me-BAFP)(Cl)]+ has taken place. Next, we used 1H NMR spectroscopy to determine the spin state of the iron center in the oxidized porphyrin complex. As previously mentioned, the β-pyrrole hydrogen atoms of the porphyrin coligand shift as a function of the iron oxidation and spin state. For the ferric chloride precursor (hs, S = 5/2), the chemical shifts range from 2 to 80 ppm and are paramagnetically broadened in deuterated dichloromethane. The broad peak at ∼80 ppm corresponds to the β-pyrrole hydrogen atoms of the porphyrin coligand (Figure S7, top), which is indicative of a hs ferric heme complex. The addition of excess [N(C6H4Br-4)3][SbCl6] to the ferric chloride complex results in a 1H NMR spectrum with chemical shifts that range from 80 to −11 ppm and that are still paramagnetically broadened, as shown in Figure S7, bottom. The upfield shift of some of the porphyrin peaks is indicative of oxidation of the porphyrin macrocycle.63,64 Analogous to the ferric chloride precursor, the β-pyrrole hydrogen atoms are observed as a broad peak at ∼83 ppm. This demonstrates that [Fe(3,5-Me-BAFP)(Cl)]+ is a hs ferric complex with an oxidized porphyrin macrocycle. It should be noted that the electronic structure of [Fe(TPP)(Cl)]+ has been studied in detail with techniques probing the magnetic properties such as NMR magnetic susceptibility, SQUID magnetization, and Mö ssbauer spectroscopy to confirm the quintet (S = 2) ground state of this complex.58,60,64,65 On the basis of these results, we can further conclude that the electronic structure of the oxidized complex [Fe(3,5-Me-BAFP)(Cl)]+ is best described as a hs ferric complex (S = 5/2) that is antiferromagnetically coupled to the porphyrin radical coligand (S = 1/2), resulting in a total spin of S = 2. In summary, we generated the ferric porphyrin radical complex [Fe(3,5-Me-BAFP)(Cl)]+ in pure form and characterized this complex by UV−vis, EPR, and 1H NMR spectroscopy, which provides an important benchmark in terms of identifying the location of the oxidation in complex 1. Formation of a Model Complex for Intermediate I. First, the chemical oxidant [DAcFc][SbF6] was synthesized by the reaction of 1,1′-diacetylferrocene with [thianthrene][SbF6]66 in dichloromethane, as previously reported by our laboratory.30 For the following chemical oxidation studies, the oxidant was dissolved in cold DME because of its poor solubility (